Pricing Research On Raroc Model Of Online Loan Based On Optimal Credit Risk Forecast Model

According to 84 online loan platforms combed by online loan houses,as of June 2019,there were still 21 online loan platform projects in default of more than 4 per cent,of which 12 had a default rate of more than 10 per cent.The default rate of online loan platform is high,mainly because of the lack of online loan industry needs a complete risk control system,and lacks effective means to identify the borrower's credit.Therefore,it is particularly important for Network loan platform to adopt appropriate technical methods to accurately evaluate and price the credit risk of borrowers.At present,big data mining technology is developing day by day,which is popular in the Internet industry.Can mainstream data mining technology be applied to the financial industry to evaluate the credit risk of online loan borrowers? Which of these data mining methods works best? In addition,the price of Online loans is also an important indicator to reflect the credit status of borrowers.Can the fixed pricing method used in Online loan platform reflect the credit risk of borrowers? Which of the fixed pricing methods is better than the RAROC method? This paper will be based on the transaction data of Renren loan platform,using Logistic,Bayesian network,C5.in this paper,we will use the transaction data of Renren loan platform.6 big data mining algorithms,such as decision tree,neural network,random forest and support vector machine,are used to evaluate the credit risk of the borrower.Then,a kind of credit evaluation model with excellent performance was selected by using the accuracy,KS value,AUC,Gini coefficient and BS value to compare the prediction effect of the analytical model with the method of 20 cross-examinations.Finally,using the credit evaluation result of the model,the loan is re-priced through the RAROC method,and compared with the original pricing price,the rationality of the original pricing method for everyone is compared.It is found that using six data mining methods can effectively evaluate the credit risk of the borrower of the network,and it can be found that when the credit rating of the borrower is increased,the default rate is obviously reduced.in which,the effect of the neural network model credit risk assessment is the best,the second is the C5.0tree,and the third is the SVM model,and the original interest rate and the RAROC pricing interest rate of the online loan are compared and analyzed by using the RAROC method,It can be found that the existing fixed interest rate pricing method for all can not reflect the matching of the loan yield and the risk,and the lack of reasonable measure and sufficient compensation for the risk of the loan.Using the RAROC pricing method,the matching of the loan yield and the risk can be realized.